From: Tom Lane <tgl@sss.pgh.pa.us>
To: Dean Rasheed <dean.a.rasheed@gmail.com>
cc: Yugo NAGATA <nagata@sraoss.co.jp>,
        PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>,
        Dave Page <dpage@pgadmin.org>
Subject: Re: Numeric x^y for negative x
In-reply-to: 
 <CAEZATCX-giGez-yYFUbBAkKhUOOG+rAQuTbnXkemJma8gouWSQ@mail.gmail.com>
References: 
 <CAEZATCW6Dvq7+3wN3tt5jLj-FyOcUgT5xNoOqce5=6Su0bCR0w@mail.gmail.com>
 <CAEZATCUoVETahg892Qu3umjXX-V8J-F2xqQqunxckd5Se8KLyg@mail.gmail.com>
 <CAEZATCWAiXp7ocU3AMV6Pyvgv3O8x4-UBY+d=w+pziXXegdyOA@mail.gmail.com>
 <20210720181509.51a037ed0a8cb0f6d3ffda3a@sraoss.co.jp>
 <CAEZATCXGk5QQPRNjeXrd7EAkzhwwjgjLhEOp=zZwktX=zYradQ@mail.gmail.com>
 <20210722141136.573c11dfc06f4dd46b661bca@sraoss.co.jp>
 <CAEZATCW+yCy_11f56oDUecQzSrdO18gwPmOVwgbw6L0_HYJJeg@mail.gmail.com>
 <CAEZATCUWUV_BP41Ob7QY12oF+qDxjTWfDpkdkcOOuojrDvOLxw@mail.gmail.com>
 <2394813.1628179479@sss.pgh.pa.us>
 <CAEZATCX-giGez-yYFUbBAkKhUOOG+rAQuTbnXkemJma8gouWSQ@mail.gmail.com>
Comments: In-reply-to Dean Rasheed <dean.a.rasheed@gmail.com>
	message dated "Thu, 05 Aug 2021 19:27:16 +0100"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <2449470.1628218684.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Thu, 05 Aug 2021 22:58:04 -0400
Message-ID: <2449471.1628218684@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/2449471.1628218684%40sss.pgh.pa.us>
Precedence: bulk

Dean Rasheed <dean.a.rasheed@gmail.com> writes:
> On Thu, 5 Aug 2021 at 17:04, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> It looks like castoroides is not happy with this patch:
>> https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=3Dcastoroides&d=
t=3D2021-08-01%2008%3A52%3A43

> Hmm, there's something very weird going on there.

Yeah.  I tried to reproduce this on the gcc compile farm's Solaris 10
machine, but the test passed fine for me.  The only obvious configuration
difference I can find is that that machine has

$ cc -V
cc: Sun C 5.10 SunOS_sparc Patch 141861-10 2012/11/07

whereas castorides' compiler seems to be a few years older.  So this
does seem like it could be a compiler bug.

			regards, tom lane